A major complication that you do not address is how to do confidence testing on the results of your pricing difference. There is an important difference between "A did better than B but who knows why" and "We're confident that there is a real difference." Figuring that out is not easy.
Yes, that's right I should have touched that point. I think the revenue for each variation can be approximated as normal curve and then a hypothesis test can be done to see if difference in mean is statistically significant.
